Solid sorbents for the recovery of nitrogen oxides (NO and NO2) for MHD exhaust gases
by Christopher John Evoniuk
A thesis submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE
in Chemical Engineering
Montana State University
© Copyright by Christopher John Evoniuk (1976)
Abstract:
Nitrogen oxides (NO and NO2) are air pollutants produced in high, temperature combustion processes.
Th generating power by the Magnetohydro dynamic (MED) process high temperatures are used and
subsequently nitrogen oxides are produced. Two methods which:' are technically feasible for the
control of the nitrogen oxides (NO ) are: combustion modification to reduce the amount of NO
produced and recovery of the NO for the manufacture of nitric acid or nitrate fertilizer. This research
was performed to study solid sorbents for the recovery of nitrogen oxides.
Results of this work indicate that the carbonates of lithium, sodium or potassium and the first row
transition metal oxides are not effective absorbents. However, mixtures of sodium or potassium
carbonate with manganese dioxide or ferric oxide are very effective absorbents of NO and NO2. Ten
grams of a 1:1 mole mixture of alkali carbonate and metal oxide in a temperature range from
380-450°C reduced the NO concentration from 9000 ppm to less than 100 ppm from a feed of
500Xml/min consisting of 15% CO2, 18% O2, 66% N2 and .9% NOx in a contact time of less than 4
seconds. Other equally effective sorbents were potassium manganate and decomposed mixtures of
sodium nitrate and ferric oxide.
Sorbents could be regenerated and then reused by heating in a temperature range of 600-700°C. During
the regeneration process, NOx could be recovered at concentration levels from 20 to 50% and the
major component of the NOx was NO2. STATEMENT OE PERMISSXOli TO COPY
In presenting t h i s thesis in, partial, fulfillment of the require­
ments for an advanced degree at Montana State University, I agree that
the Library shall make it freely available for inspection,
I further
agree that permission for extensive copying of this thesis for schol­
arly purposes m a y he granted b y m y major professor, or, in his absence
b y the Director of Libraries.
It is understood that any copying or
publication of this thesis for financial gain .shall not be allowed
without m y written permission.
Signature'
Dat e
:
Vf 7^1
SOLID S ORBEMTS FOR THE RECOVERY OF
' MITROGEM OXIDES (MO AMD NO )
FROM MHD EXHAUST GASES ^
by
CHRISTOPHER JOHN EVOMIUH
A thesis submitfed in partial fulfillment
of the requirements for the degree
of
MASTER OF SCIENCE
in
Chemical Engineering
Approved:
Chairperson, Graduate Committee
Head, Major Depantmerrt''"]
Graduate 9Dean
MONTANA STATE UNIVERSITY
Bozeman, Montana
Decemher, 1976
tit
ACKUOELEDGEMENTS
TIie author wishes to thank!.the faculty and staff of the Depart­
ment of CIiemical Engineering for tIieir assistance, especially Dr. F.P.
McCandless for his Iielp and suggestions.
Special thanks is extended'
to the aut h o r 1s wife, Marian, for her help in typing and editing the
thesis and for Iier patience.
Financial support for this study was provided b y the Energy
Research and Development Administration for Montana State University
M H D Program Contract Mo. E (.49-18)-l8ll. Task HIV, administered
through the Montana Energy and MHD Research and Development Institute
Inc.
TABLE OF CONTENTS
Page
T O T A .......
ACKNOWLEDGEMENTS
Ii
..................
LIST OF TABLES ........___ ............... ................ .
IiI
v
'
LIST OF FIGURES ............
vi
ABSTRACT
IX
INTRODUCTION
I
OBJECTIVES .... ................ ■--- ........---- ■............ .
6
APPARATUS AND P R O C E D U R E .......
7
RESULTS
.......
13
APPLICATION .............................
67
CONCLUSION ......................
70
RECOMMENDATION .............
72
LIST OF R E F E R E N C E S ... ...........'.......... ....................
73
V
LIST OF TABLES
Table
I.
Page
Equilibrium Constants and Rate Constants
.................l4
II.
Equilibrium Values
of NO and NO0 ..........
15
■III.
Equilibrium Values
of N^O^ ...........
l6
IV.
Equilibrium Values
of N g O ^ ..... .
l6
V.
VI.
NO^ Absorption b y the Metal Oxides
..... .
20
NOx Absorption b y Mixtures of Na^CO^ and
Metal Oxides ............................
VII.
24
Effect of Flow on Temperature Gradients
in the Reactor and NOx Absorption ...... ......... .
VIII.
33
NOx Absorption b y KMnO^ and Mixtures of N a N O ^ 5 Na^CO^,
KgCO^j KNO^ with Fe 5O^ and MhO^ ..:.......... ...... .
IX.
35
Equilibrium Constants for Reactions
Involving NO and COg .....................................
X.
Melting and Boiling Points of Alkali
Carbonates and Oxides
XI.
XII.
XIII.
49
....................................
Thermal Stability Characteristics of Alkali Oxides
Possible Reactions of the Sorbents
53
.... 54
.......... ..........
58
Melting and Decomposition Temperatuhes of
Alkali Nitrates, Nitrites and Hydroxides ............... 60
XIV.
XV
Cost of 7.28 x 10^ Moles of Sorbent Mixture
Cost of Chemicals
67
69
LIST OF .FIGURES
Figure
1.
FI ott Diagram of Apparatus
2.
Reactor Cross Section ...
3.
NO
4.
NO
5.
NO
X
X
X
6 . NO
7.
NO
X
X
8 . NO
9.
NO
10.
NO
ll.
NO
X
x.
X
12 . NO
NO
l4.
NO
15.
NO
16.
NO
17.
NO
CO
NO
H
13.
19.
NO
20. NO
X
X
X
X
X
X
X
X
X
X
Absorption Curve for Li 2CO 3
Absorption Curve for Na2CO3
Absorption Curve for KgCO 3
Absorption Curve for TiO 2 .
Absorption Curve for Co2 Os
Absorption Curve for MnO 2 .
Absorption Curve for Pe2 Os
Absorption Curve for PbO 2 .
Absorption Curve for Al2O3
Absorption Curve for Na2CO
+ TiO 2
Absorption Curve for .Na2CO3 +
Absorption Curve for Na2CO3 + Cr2 O
Absorption Curve for Na 2CO 3 + MnO 2
Absorption Curve for Na 2CO 3 + Fe2O
Absorption Curve for Na 2CO 3 + Co2O
Absorption Curve for Na2CO3 + NiO
Absorption Curve for Na 2CO 3 + CuO
Absorption Curve for Na2CO3 + Cu2O
Vii
Figure
Rage
21.
Absorption Curve for NagCO^ + Al^O^ ..................
29
22.
MOx Absorption Curve for Decomposed MaWO^ + Fe^O^ .......
31
23.
MO^ Absorption Curve for Ma^ Fe^ O ^ .... .............
31
2i.
Effect of Temperature on MOx Absorption b y
i
25.
Decomposed MaMO^ + Fe^Og ......... ........ ................
Effect of COg on MOx Absorption b y Decomposed
MaMO
.26.
32
+ Fe 2O 3 ......... .............. ....... . ..... ........
34
Effect of MO to Og Ratio on MOx Absorption b y
Decomposed M a M O 3 + FegQ 3 ......... ........................ . 34
27.
MOx Absorption Curve for MagCO 3 + FegO 3 (Bed)
...........
37
28.
MOx Absorption Curve for MaMO 3 + FpgO 3 (Bed) .............
37
29.
MOx Absorption Curve for KgCO 3 + FegO^ (Bed) ............
38
30.
MOx Absorption Curve for Decomposed KMO 3 + FegO^ (Bed) .. 38
31.
MO
(Bed) ............
39
32.
MOx Absorption Curve for KgCO^'+ MnOg (Bed) .............
39
33.
MOx Absorption Curve for KMhO^ + SiOg (Bed) .............
4i
3^.
MOx Absorption Curve for Decomposed MaMO 3 + MnOg (Bed) .. 4i
35.
MOx Absorption Curve for MagCO 3 + SiOg (Bed)
............
43
36.
MOx Absorption Curve for .KgCO3 + SiOg (.Bed) .............
43
37«
MOx Absorption Curve for MnOg + SiOg (.Bed) ............... • 44
38.
MOx Concentration During Regeneration of KgCO 3 + MnOg ... 45
39.
Regeneration of KgCO 3 + MnOg ..............................
x„
Absorption Curve for Ma CO
2
3
+ MnO
^
45
viii
Figure
Page
40.
Loading Curve for KgCO^ + F e g O ^ .......... •................
4-7
41.
Loading Curve for KgCO^ + MnOg ........... ................
4%
■42.
Effect of Metal Oxides on Decomposition of Carbonates ... 56
43.
Amount of
.... .....
63
44.
Amount of Og Released From KgCO^ + FegO^ ........ ........
63
45.
Absorption of Og b y KgCO^ + MnOg and KgCO 0 + FegO^ ......
66
Or
. Released From K d
0CO Z
0> + M nO d
0
td
ix
ABSTRACT
Nitrogen oxides- (NO and NO^ I are air pollutants produced in high,
temperature combustion p r o cesses. T h generating power b y the Magneto­
hydro dynamic (MED) process high, temperatures are used and subsequently
nitrogen oxides are produced.
TVo methods which:' are technically feas­
ible for the control of the nitrogen oxides (NO ) are:
combustion
modification to reduce the amount of NO produced and recovery of the
NO
for the manufacture of nitric acid or nitrate fertilizer.
This
research was performed to study solid sorbents for the recovery of
nitrogen oxides.
Results of this w o r k indicate that the carbonates of lithium,
sodium or potassium and the first row transition metal oxides are not
effective absorbents.
H o w e v e r , mixtures of sodium or potassium car­
bonate with manganese dioxide or ferric oxide are very effective ab­
sorbents of NO and NO^.
Ten grams of a 1:1 mole mixture of alkali
carbonate and metal oxide in a temperature range from SBO-U^O0C red­
uced the NO concentration from 9000 ppm to less than 100 ppm from a
feed of 500Xml/min consisting of 15% C O g , 18% O g , 66% N^ and .9% NO^
in a contact time of less than 4 seconds.
Other equally effective "
sorbents were potassium manganate and decomposed mixtures of sodium
nitrate and ferric oxide.
Sorbents could be regenerated and then reused by heating in a
temperature range of '600-T00°C.
During the regeneration process, NO^
could be recovered at concentration levels from 20 to 50% and the
major component of the NO^. was N O g .
INTRODUCTION
Removal and control of nitric oxide from exhaust gases is a prim­
ary concern for all combustion processes. - Nitric oxide (NO) is not
.
considered as an irritant in itself and is not considered to have ad­
verse health effects at concentrations found in the ambient atmosphere
Its greatest toxic potential is due to its oxidation to nitrogen di­
oxide (NOg).
Nitrogen dioxide exerts a toxic effect on the lungs
and concentrations of greater than 100 ppm are lethal to most animal
s p e c i e s , including man. ■ Long term exposure to low concentrations does
not cause acute inflammatory responses but does have an accumulative
sustained effect.
Results of community exposure to NO^ levels of .062
to .109 ppm for long periods of time (6 months) gave an increased
respiratory illness rate of 18% (Pollution Control Technology, 1973).
On a global basis the total amount of nitrogen dioxide generated
I
by n a t u r e .exceeds that made by man-made technological sources.
Nat­
ural scavenging processes in nonurban areas keep background levels on
the order of 2 -■ 4 ppb.
In urban areas the levels are frequently
higher because pollutants are added faster than scavenging processes
can control them.
Fuel combustion is a major source of nitrogen dioxide in air
pollution.
Combustion of coal, oil and gasoline accounts for the ma j ­
ority of man-made emissions.
The estimated global breakdown of nitro­
gen dioxide emissions in the year
1965 is as-follows (Seinfield, 1975)
2
Source
Emissions in Trillion grams of
Coal combustion
2h.h x 10
Petroleum processes and combustion
20.2
"
Natural gas combustion
1.9
«
Miscellaneous
1.5
" •
(fires, etc.)
EPA emission standards for KOg from coal fired steam generators
a r e -0.TO Id. per million Btu of heat input or "approximately 600 ppm
(Walters and G o o d w i n , 197^)•
an efficiency of
would be
For a 1000 Megawatt MHD generator having
50% it is estimated that the exhaust gas flow rate
1.25 million standard cubic feet per minute and could contain
from less than 600 ppm NO to greater than 30,000 ppm NO
(3%) depen- .
ding on the combustion process used (Hals and Lewis, 1973).
To comply with EPA standards the MHD process would have to limit
NOx (NO and NO^) emissions.
Of the methods which appear promising,
the most practical would be either combustion modification or absorp­
tion of the NOx produced.
Previous Work
Combustion modification would reduce the amount of NO formed by
lowering of the combustion temperature (Pepper, Eustic and K r uger,
1972).
However, a preliminary economic" analysis indicated that if the
nitric oxide is collected and converted to nitric acid or a nitrate
fertilizer, the income from marketing the product would be much greater
3
than that from the sale of electricity produced, b y the MHD generator
(Wright and Yo u d e r i o n , 1976; Hals and Jackson,
1969).
Since profit­
ability is the prime concern of any industrial process,collection of
the nitric oxide would be preferred over combustion modification.
A method which is technically feasible for the recovery of the
nitric oxide is absorption
by liquids or solids.
Liquid absorption
appears promising for it can be used to absorb both HO and SO^. Usually
alkaline solutions or sulfuric acid are used in conjunction with equi­
molar concentrations of HO and HO^ (Hals and Jackson,
1969).
Absorp­
tion of HO^ by liquid systems is' improved by having equimolar concen­
trations of the two oxides rather than just HO^ (Koval and Roberts,
i 960).
The increased absorption is due to the formation of HgO^
from the HO and HO^, whi c h is then absorbed to form nitrous acid.
Hitrogen dioxide would have to be recycled to the incoming gas stream
to achieve the equimolar concentration.
Among the alkaline scrubbing systems which have been tested,
lime-water or magnesium hydroxide solutions have been reported as
showing the most promise (Bartok, et al, 1971).
The magnesium hydrox­
ide is said to be better due to advantages in regeneration of the nit­
ric oxide from the nitrite formed.
From the standpoint of forming a nitrate fertilizer, liquid ab­
sorbents form nitrites which would have to be oxidized to nitrates.
Also the regeneration of liquid sorbents produces HO and little HOg
k
which, means that an added investment would he needed to oxidize the
W
to N O .
-Among solid sorbents, metal oxides, especially manganese and fer­
ric oxides, have been reported to show technical potential (Leung and
G i d a s p o w , 1974).
Patent literature indicates that sodium and calcium
carbonate or calcium oxide can be used to. absorb
M o r e l l o , and Peters,
1969).
but not NO (H a r r e s ,
Nitric oxide is removed at normal temper­
atures and pressures by inorganic salts of hydroxy acids
(FeSO^, KMnO^,
KClO^) by inorganic oxides (Na^O^, PbO^) or by inorganic halides (KI,
CuCl2 ) (Lewis, 1975).
A ferric oxide-sodium oxide sorbent developed by R» A. Ogg and
J. D. Ray (U.S. Patent 2,684,283) absorbs NO at 400°C forming sodium
nitrate and little sodium nitrite.
The sorbent was prepared by heat­
ing 2 moles of sodium nitrate (NaNO^)-and I mole of ferric oxide
(Fe 2O ^ ) in a temperature range of JOO to 900°C until the sodium nit­
rate was largely decomposed and no nitrogen oxides were being given
off.
When a gaseous mixture consisting of 0.7% NO, oxygen and nit­
rogen was contacted wi t h the sorbent the NO was absorbed and sodium
nitrate was formed.
Nitrogen oxides could then be recovered in a con­
centrated form on heating the sorbent to a temperature of over 700°C.
A sorbent of this type would be particularly useful since a nitrate
fertilizer could be directly formed or a concentrated stream of
nitrogen oxides could be produced for the manufacture of nitric acid.
5
The importance of HO and.NO^ concentrations in the formation of nitric
acid is best shown hy the following example:
I atm and 25°C containing 5$ H
of which
For a gas mixture at
Q0% is oxidized to NOg the
equilibrium- strength of nitric acid with this gas mixture is
weight
(C h i lton„
55% by
1968), while if the gas contains 50% HO^ of which
80% is oxidized to HOg then the equilibrium strength of nitric acid is
rJ0% b y weight.
The increased concentration is useful if the nitric
acid is to be concentrated.
Also by increasing the pressure of the
concentrated NO^ stream the equilibrium strength of nitric acid
would be well over
70%.
OBJECTIVES
The purpose of this research
was to develop a solid for the re­
moval of NO and NO^ (NO^I from the M HD exhaust gases and either convert
them to a nitrate or on regeneration produce a more concentrated
stream of NO^ or NO.
Other objectives were to determine the factors
w h i c h affect the absorption and regeneration processes.
APPARATUS AED PROCEDURE
Figure I is a schematic of the apparatus used.
Individual gas '
■
streams were passed through rotameters and into a mixing chamber to
r
form a test gas of similar composition to that expected for the MHD
exha u s t .
From the mixing chamber the gas could be passed directly to
the HO^ analyzer to determine the NO^ composition.or it could be
passed to the reactor and then to the NO^ analyzer.
Septum ports were
located on b o t h sides of the reactor to permit taking of gas samples
for analysis of CCR,, E^ and CR,.
Figure 2 is a schematic of the reactor.
a stainless steel pipe
The reactor consisted of
10.5 inches long with an outside diameter of
7/8 inch and an inside diameter of 5/8 inch.
couple well made from l A
A stainless steel thermo­
inch O.D. and 5/32 I.D. tubing was situated
in the center of the reactor.
Approximately 7 inches from the top of
the reactor a stainless steel porous disk was placed.
Beneath the
disk the reactor was packed with stainless steel rings which served to
preheat the gas coming in.
of the disk.
Materials to be tested were placed on top
Glass wool packing at the exit of the reactor prevented
solid particles from blowing o u t .
After each run the reactor was
cleaned with dilute hydrochloric acid, rinsed with distilled water
and dried.
The reactor was heated by an inconel pipe wrapped with 6.5 feet
of ceramic beaded n i chrome wire (.9 ohm/ft) and insulated b y a 6.5 inch
diameter housing packed with zeolite.
Rotatmeters
Air or
N O 2 or NO
Mixing chamber
Sample port
Exhausts
Water cooler
Analyzer
Dilutent
rotatmeters
Filter
Reactor
Temperature
recorder
Figure I.
Flow Diagram of Apparatus
Sample port
9
To analyzer
Inconel pipe
•Glass wool
Thermocouple
well
Heater
coils
Absorbent
Insulation
Porous
stainless steel
disk
Housing
Stainless steel
Stainless steel
rings
Glass wool
Feed gas
Thermocouples
Figure 2
Reactor Cross-section
10
Sorbents
The sodium oxide-ferric oxide sorbent was prepared by heating a
2:1 mole ratio mixture of sodium nitrate and ferric oxide at red heat
in a small rotating kiln over a large burner for approximately one
hour.
Gravimetric analysis indicated a weight loss of approximately
31% w h i c h corresponded to a formula of NagOFe^O^ for the remaining
material.
Bulk density of the substance when loosely packed was. beO
tween
1.33 and 1.44 g/cm .
Transition metal oxides and the alkali carbonates were tested in
the form that they were obtained from reagent grade chemicals.
tures of the oxides and carbonates were made on a
Mix­
1:1 mole basis and
10 g. of the resulting mixture was used for testing.
The chemicals
were dry mixed with a mortar and pestle.
Gas Mixtures
Standards for calibrating the analyzer were made b y diluting .50%
HO (5 0 0 0 ppm) with room air.
The reactant gas mixtures also contained
HO diluted w i t h room air to produce the desired HOx concentration,
generally between .75% (7 5 0 0 ppm) and 1% ( 1 0 ,0 0 0 p p m ) .
Hitric oxide
reacted with oxygen in the air to produce a final gas mixture con­
taining approximately .5% HO and .5% H O ^ .
Carbon dioxide was added to the feed stream and the resulting
concentration of COg was approximately 15%.
Due to plugging of the
11
analyzer B y high, concentrations of moisture, no appreciable amount of
HgO was included in the feed gases other than what was in the room air.
To determine the effect water had on the absorption, the air was
saturated prior to being mixed w i t h HO at 2It0C to get a moisture con­
tent of approximately
2 - 3$.
Analysis of HO and HOg was accomplished using a Thermo-Electron
Model IOA Self-Contained Chemiluminescent HO-HO^. Gas Analyzer, which
has the capability of measuring from 0-10,000 ppm of HO or H O g .
ni­
trogen and oxygen analysis was performed on a Beckman gas chromatograph
using a column packed w i t h molecular sieve- type 13-X.
Carbon dioxide
analysis was performed on a Varian gas chromatograph using a column
packed w ith Porapak Q-S (Waters Associates).
Hitrate Analysis
Hitrate analysis was performed b y using an Orion specific ion
electrode model 93-07.
The solid sorbent was dissolved in water and
the p H of the resulting solution was adjusted to approximately 5.
The resulting solutions were then compared to known mixtures contain­
ing nitrate and nitrite ions.
Presence of the nitrite ions was det­
ermined in the following m a n n e r ;
The amount of HO^ absorbed by the
sorbent was determined from the absorption versus time curves.
Stand­
ard solutions of nitrate and nitrite ions -were made to approximate the
concentration anticipated when the. sorbent w a s .dissolved in water.
12
Comparison of the reading for the sorbent solution and for each stand­
ard solution tras then made to determine if nitrite or nitrate ions
were predominately present.
RESULTS
Due to the reaction of NO with.
to form NO2 , it was necessary
to determine the actual compos'ition of the gas entering the reactor.
Writing the chemical equation for the oxidation of NO to NO 2 as:
'2N0 i O2 = 2N0g
the rate expression for the formation of N O 2 'can he written as:
dpNOgyzdt =
NO P0g
where,
" rate of change in the partial pressure of NOg
dPN0
k = reaction rate constant
PN0 = partial pressure of NO
Pq
= partial pressure of Og
The rate expression can be written in a more useful form L y letting
a = initial partial pressure of NO (atm).
■ L = initial partial pressure of Og (atm)
c = partial pressure of NOg at time t
t = time from start of reaction (sec.)
k = reaction rate constant:
' and.T is in pK' (Chilton,
k = exp( 6^l/T - 0,72p)
(atm ^sec 1 )
1968) .
substituting these variables in the original rate equation gives the
expression:
dc/dt = k(a T c)2 (b - c/ 2 )
Xf b is m u c h larger than c /2 then the last term on the right hand side
Ill
reduces to "b,.
TKe resulting expression can be easily integrated and
solved for c to give;
c = a(l - I / CaKbt t ill
If a = .01, Tb = .21, k g .38 and t = 27 then
c =
Xf a =
.0068 atm C6800 ppm I
.008 and the other variables remain the same then
c =
.00506 atm (5060 ppm]
Approximately 50% of the original NO is' oxidized to UO^ in the time
(2.
7? s e c ) it takes, the feed, stream to reach the reactor.
To determine if the NO^ decomposes significantly as it passes
through the preheat section, the equilibrium constant (K ) and the
P
rate constants for the forward (k^] and reverse (k^] reactions need
to be known.
Table X lists values of k^ calculated from the .expression;
1V M d = kP = P N 0g/P NOp O 2
TABLE I;
°C
38
K
I X
Equilibrium Constants and Rate Constants
200
5
H 6
10
IlOO
I x IOp
I x
2.2 x IO 1
li.3
1.69
2.2 x
ir;3 X 1 0 - 5
•
10
I
(Chilton,
600
800
-I
I x
1968)
10
-3
1.7 x 10 0
P
V
kd
IO-10
,.169
1 .02.
10 . RJ
.Jh5
7.'ii5 x IO 2
As can be seen from the table the decomposition reaction does not
proceed at a significant rate until the temperature exceeds
1|00°C.
15
Since the. decomposition reaction vlll Be Important aBore UOO 0C , it is
important to know* vzhat the' equilibrium .concentrations' of NO and NO^
will he above this- temperature.
Table H
lists- some w a lues for the
concentrations- of NO and NO^ when the initial amount of NO added to a
stream containing
TABLE II:
21$ Og is- known.
Equilibrium "Values- of NO and NOg
Initial NO concentration =
10,000 ppm
®c
Uoo
600
800
N O g (ppm)
677U
17U 0 -
U
NO(pprn).
3226
8260
9996
Initial concentration = 8,000 ppm
N O g (ppm)
5U 20
1390
3
N O (ppm)
2580
6610
7997
Comparison of these values w i t h the values calculated for the feed
composition indicates that the concentration of NO and NOg entering
the reactor will be close to the equilibrium concentration of NO and
NOg at U-OO0C,
Other nitrogen oxide species that could form from the
mixture of NO and NOg are dinitrogen trioxide (NgO^ I and dinitrogen
tetroxide (NgO^).
and Table IV.
Equilibrium values for these are shown in Table III
1
16
TABLE III:
Equilibrium Values for
(Chilton, 1968)
K = P
/PT
P
N 2O 37 NO NO 2
Reaction NO + NO 2 = NgO
0.272
0 .4 7 5
K
0 .1 4 5
Equilibrium Values for N 5O^ (Chilton,
TABLE IV::
Reaction
45
35
25
°C
2N 02 = N 2O 4
kp
°C
P
20
65
12
=
V ZV
1 1
2
80 -
60
Uo
2.7
O
CVJ
K
0
1968)
.'67
Inspection of the equilibrium values indicates that neither of these
species would be .-significant" at temperatures above I OO 0C.
•.Graphs and Order of P r e s e n t a t i o n -•
•
-
•
Some points which are important in reading of the graphs are as
follows:
The feed concentration is in ppm of NO while the exit conc­
entration'' is in ppm of NO^ (NO + NO^).
It should also be noted that
desorption is occurring when the exit concentration (NOx ) exceeds the
original NO concentration in the feed.
Since absorption tests were
performed with changing temperature, the concentration of NO
in the
x
gas stream leaving the reactor and the corresponding temperatures are
17
plotted against time.
The original order of testing sorbents was to examine the sodium
oxide-ferric oxide sorbent in detail since it was reported to be a very
effective solid sorbent.
Then mixtures of alkali carbonates and metal
oxides were tested as was.the individual components of the mixtures to
determine if a more effective sorbent could be found.
For unification
the results are presented in the order that follows.
UOx Absorption.by the Alkali.Carbonates
The concentration of UOx leaving the reactor and temperature
curves for runs made with
10 grams of Li^CO^, Ua^CO^, and K^CO^
are shown in Figure 3, I and 5.
Each carbonate has an initial absorp­
tion m aximum at approximately IOO0C and at 350 - U50°C..
The absorp­
tion at I O O 0C does not occur after the carbonates have been heated.
Absorption capacity increases from lithium to potassium.
carbonate is unique in that it reduces the UO
(UO + UO
Potassium
) concentra-
tion leaving the reactor below the UO^ concentration entering the
reactor at temperatures between 350 and U^O 0C.
It should be remem­
bered that of the original UO that is injected into the feed stream,
approximately half is oxidized to UO^ before reaching the reactor.
Also as UOg is absorbed the possibility exists that the rate of UO
oxidation will increase creating more HOg which can be absorbed.
'- k
18
too ^
£
Sxlt Concentration
500
3
I
kOO
300
Figure 3.
NO
x
Absorption Curve for Li0CO
2
3
Sa2CO3
Feed - $00 al/mio air
(ppm NOx x IO--3)
9000 ppm NO
___
Temperature
Minutes
Figure 4.
NO Absorption Curve for Na0CO0
x
2 3
I
19
[,CO,
Peed - 500 al/ain air
9000 ppm 30
_
_ Temperature
Temperal
_
1,300 ppm
— 300 ppm
Mlnutea
Figure
5.
N O ^
Absorption Curve for
K g C O ^
20
A b s o r p t i o n .of NO^ .by .Metal.Oxides
6 through- 11 are the results of the runs made with metal
Figures
oxides TiO2 , M n O 2 , FegO^, CogO^, PbOg and AlgO^ (activated alumina).
Of these oxides, alumina and lead dioxide showed the greatest potential
for absorbing N O ^ . ■
At room temperature AlgO^ absorbs NOg and possibly some N O . (Fig­
ure .11) while PbOg absorbs primarily NOg (Figure 10).
Both oxides have
an absorption maximum at about 300 - U00°C, but AlgO^ desorbs in the
temperature region from 20°C to 300°C.
Above a temperature of U00°C
AlgO^ and PbOg desorb NO^ and after the desorption AlgOg will absorb
more NO^ at U 00°C than previously
(9200 ppm compared to 800 p p m ) , while
PbOg absorbs less (3500 ppm compared to 6250-ppm).
Manganese oxide absorbs mostly NOg at room temperature but as the
temperature increases from 26°C to 500°C desorption occurs (Figure
8 ),
then as the temperature decreases from 500°C to 300°C some absorption
occurs again.
TABLE V.
NO
x
'Table 5 summarizes the performance of the metal oxides.
Absorption by the Metal Oxides
■■ To (ppm)
Oxide
From (ppm)
TiO 2
.9,000
.7,000
MnOg
10,000
7,500
PbOg
7 *000..
750
From(ppm)
To(ppm)
COgOg
9,000
" 6,800
Fe2 Og
10,000
8,900
10,000
800
"Oxide
M
2O 3 '
21
TOO
600 u
8
U00 I
300
6 . NO^ Absorption Curve for TiO^
600
500
Uoo
300
200
100
Figure 7.
NO
Absorption Curve for Co0O
2 3
I
Temperature
Figure
22
-T
•o
700
500
I
I
IiOO
Figure
Tempera!
600 F
I
8 . NOx Absorption Curve for MnO^
Fe2O3
Feed - 500 ml/min air
8,900 PJM
10000 prpm NO
_ Temperatrue
Temperature (°C)
------- ,0x
Minutes
Figure 9.
NOx Absorption Curve for Fe^O^
23
Feed - 500 al/min air
Tempei
7000 rpo SO
Minutes
Figure 10.
NO^ Absorption Curve for PbO^
Activated alumina (AlgO^)
Feed - 500 ml/min air
10000 ppm HO
I
Figure 11
NO^ Absorption Curve for Al^O^
2b
It should be noted that these absorption.tests were performed when .
no COg was present in the feed.
Absorption b y Mixtures of Metal Oxides and Alkali Carbonates
Results of the runs made with mixtures of the first row trans­
ition m etal oxides and sodium carbonate are shown in Figures 12 to
21.
In all cases a maximum absorption occurs in the temperature
range from 350 ' to
h50°C.
No increase in absorption of HO^ occurs
over that for' R a ^ C O ^ .until, oxides of manganese^,.-iron,..c o b a l t , n i c k e l ,
or copper are used.
The oxides of manganese, iron, nickel, and cobalt
when mixed with sodium carbonate reduce the
the reactor to less than 100 ppm.
run made w i t h M n O g ).
concentration leaving
(Rote that CO^ was present in the
Table VI summarizes the performance of mixtures
of RagClg with metal oxides.
TABLE VI.
ROx Absorption by Mixtures of RagCOg and Metal Oxides
Oxide
I t o m (ppm)
To(ppm)
Oxide
FromCppm)
To (ppm)
10,000
3,000
COgOg
9,000
100
. 8,000
1,900
RiO
8,000
30
CtgOg
8,000
3,200
CuO
8,000
800
MnOg
8,000
50
CUgO
8,000
550
FegOg
7,500
25
AlgOg
10,000
600
TiOg
V 2 05
•
25
Ha2CO3 + TlO2
Feed - 500 ml/min air
10000 ppm NO
_
_
_ Tatrperature
7OC
600
HOO
Temiperi
500
30C
3,000 ppm
200
100
0
Figure 12.
NO^ Absorption Curve for Na^CO^ + TiOg
TOO
600
o’
,00
I
I
itoo E
Figure 13 .
NO^ Absorption Curve for Na2CO3 + VgO^
26
Temper!
Exit Concentration (ppm HO
Heed - 500 ml ,'Bin air
3,200 ppn
Minutes
Figure l4.
NO
x
Absorption Curve for Na 0CO 0 + Cr0O
2 3
2 3
Feed - U-OO al/ain air
BCCO ppr RC
Concentration (ppm
Temperature (0C)
x 10*J )
ICC ai /Bin CC
50 ppm
Minutes
Figure 15
NOx Absorption Curve for Na2CO3 + MnO,.
27
'rOO
60C v
500
I
I
too I
30C
2OC
100
Figure l 6 .
NO^ Absorption Curve for Na^CO^ + Fe^O^
600
500
p
§
300
Figure 17.
NO^ Absorption Curve for Na2CO^ + Co2O^
Tempei
8
Uoo
28
Sa2OO3 ♦ SiO
Peed - 500 al/min air
8000 ran SO
_
_
_ Temperature
------
6oo
z
200
200 ppm
Minutes
Figure l 8 .
NO^ Absorption Curve for Na^CO 3 + NiO
Sa0CO, + CuO
500 ni/ain air
8000 ppm NO
600 u
^
Tem pe:
500
800 ppm
Minutes
Figure 19 . NOx Absorption Curve for NagCO3 + CuO
29
700
600 u
Uoo
4»
g
Temperat
50 0
300
200
Figure 20.
NO^ Absorption Curve for Na^CO^ + Cu^O
700
50 0
I
S
g.
Uoo g
Figure 21.
NOx Absorption Curve for Na^CO^ + AlgO^
30
ITO^ Absorption "by Sodium Oxide and Ferric Oxide Mixtures .(Ferrates)
A mixture of sodium nitrate and ferric oxide was heated to decom­
pose the nitrate and form a sodium oxide-ferric oxide m i x t u r e .
Also
a sample of commercial sodium ferrate (WagFe^O^) was obtained and test­
ed.
Even though b o t h substances had the same molecular formula, the
commercial ferrate was red while the ferrate made from nitrate was
yellow.
(Note:
When pure NaWO^ was decomposed, a yellow substance
was produced, presumably sodium peroxide
(Wa^Og)).
The results of runs made with these two ferrates are shown in
Figure 22 and 23.
Both absorb WO^ at low temperature and also absorb
NO and WO^ at approximately 380 - 500°C, reducing the exit concentra­
tion of NO to less than 100 ppm.
test).
(CO^ is" present in the NaNO^ + Fe^O^
It is interesting to note that the temperature where absorp­
tion starts to occur is also the temperature at which the sodium nit­
rate starts to decompose and give off NO (380°C).
The extended temperature range in whi c h absorption occurs may be
due to temperature gradients in the reactor (see Table V I I ).
It can
be assumed that the middle of the reactor is where absorption first ■
occurs and also the point at which the temperature first reaches 380°C.
A test on the effect temperature had on absorption is shown in Figure
2h.
In this test no CO^ was present.
off at approximately 380°C.
Absorption of NO^ starts to drop
31
Feed - UOO ml/min air
7000 ppm NO
100 ml/min C0„
_
_
_ Temperature
------- "°x
100 ppm
Minutes
Figure 22.
NO^ Absorption Curve for Decomposed NaNO^ + Fe^O^
* V e2°i.
Feed - 500 ml/min air
10000 ppm NO
_ Temperature
,
u
60 0
°
500
I
Uoo I
300
100
50 ppm
b
Minutes
Figure 23 .
NO^ Absorption Curve for Na^Fe^O^
32
___ Taapermt-.ire
Peed —
2k0ml/mln
7500 M » 10 (.75%)
10% O2
CoocentrstIon of HO. (pj») In Ellt St
Haemlnder Ig
Temperal
-ItOO
Figure
t
2k.
Effect of Temperature on NO Absorption
by Decomposed NaNO^ + F e ^ O ^
33
TABLE V X I .
Effect of Flow on Temperature Gradients in the Reactor and
NOx Absorption
Flow (ml/min)
600
800
1100
1200
T (0C)
457
457
457
457
Tt (0C)
402
400
395
395
Tm( 0 C)
445
445
4oo
. 435
Tb (0C)
330
330
320
310
20
16
18
20
W
NOx (ppm)
T = temperature of outside wall of reactor across from T
w
m
Tj. = temperature I inch below the top of reactor
Tm = temperature I inch above the porous disk (middle of reactor)
Tj3 = temperature
.5 inch below the porous disk
NOx = concentration of exit stream (feed is 5000 ppm NO)
(Temperatures were measured by using thermocouples). *
Also in Table VII the exit concentration of NO
leaving the
%
reactor at the various flow rates are listed.
Even though the flow
rate was doubled the absorption of NOx was not affected.
Figures... 2.5. and 26 show 'the effect CO^ and O^ had on the absorp­
tion of NO and NO
*
oxide.
C.
at 400°C b y the decomposed sodium nitrate-ferric
The presence of carbon dioxide in amounts of 20% or greater
affected the absorption of N0x ; and when the NO to O^ ratio was below
1:2 the absorption of NOx was also severely reduced.
25
Decomposed SaHO^ ♦ Fe^O^
Feed - 325 ml/min air
Exit Concentration Cppe *0^
20
-
15000 ppm BO
15 -
Effect of CO
on NO
Absorption by Decomposed NaNO
+ Fe 0
Concentration (ppn NO^ x
.
1:3
I:'
Ratio of BO:Og
Figure 26.
Effect of NO to Og Ratio on NO^ Absorption by Decomposed
NaNO3 + Fe2O3
35
Performance of Sodium and Potassium Carbonates and Nitrates with
Manganese and Iron .Oxides'.........................
Figures 27 through 32 are the results of tests made with dif­
ferent combinations of manganese and ferric oxides w i t h the carbonate
and nitrates of potassium and sodium.
Table V Z I I ).
and sorbent
(The results are summarized in
In these tests equal volumes of 20-30, mesh. Ottawa sand
(.025 moles) were mixed together and poured into the reac­
tor to form a packed bed.
The volume between the bed and top of reac­
tor was then filled with Ottawa sand to keep the sorbent from separ­
ating out.
TABLE VIII.
NOx Absorption by KMnO^ and Mixtures of NaNO^, Na 2CO^,
KgCO^, ENO^ with FegOg and MnOg
NO
Sorbent
Concentration'Reduced
x
From 9000 ppm To (ppm)
NagCOg + FegOg
200
NaNOg + FegOg
150
KgCOg + Fe2 Og
25
KNOg + Fe2 Og
250
N a 2COg + MnOg
150
KgCOg + MnOg
50
NaNOg + M n O g
2500
KMnOli
25
36
Sodium nitrate-ferric oxide mixtures absorbed slightly more NO
than did the decomposed sodium carbonate-iron oxide.
The first absorp­
tion at 380°C reduced the NO^ concentration to 1900 ppm for the car­
bonate and
ppm for the nitrate, while NO
at the second absorption
x
at 380°C was reduced to 200 ppm and 150 ppm by the carbonate and nit­
rate respectively.
It was found that upon successive regeneration,
the carbonate mixture steadily improves and surpasses the performance
of the nitrate mixture.
Similarly with the potassium carbonate and decomposed nitrate m ix­
tures w i t h ferric oxide (Figures 29 and 30), the carbonate performed
m u c h better than the nitrate, reducing the NO^ concentration to less
than 50 ppm, compared to 400 ppm for the nitrate.
A few factors
worth noting about the potassium carbonate-ferric oxide mixture a r e :
(i) it absorbs more than Na^CO^ + Fe^O^ or decomposed NaNO^ + Fe^O^,
(ii) absorption starts at a lower temperature (300oC) compared to 380°C
for the sodium carbonate,
(iii) a higher temperature is needed for
regeneration (this m a y be caused by the temperature gradient in the
reactor), and (iv) more NO^ is absorbed at lower temperature than for
the previous mixtures.
The effect of using manganese dioxide with sodium and potassium
carbonate is shown in Figures 31 and 32.
Absorption was improved for
the sodium carbonate (the NO^ concentration dropped to less than
100
ppm) and there was a slight improvement for the potassium carbonate
37
TOO
600 t.
E
5 00
5
1.00
£
I
Figure 27.
NO
x
Absorption Curve for Na CO
3
+ Fe 0
(bed)
<- 3
Decomposed IiaHO3 + Fe,
Feed - UOO ml/min air
9000 ppm NO
15.8$ CO2
_
_
_ Temperature
TOO
600
500
UOO
S
I
300
200
100
0
Figure 28.
NO^ Absorption Curve for NaNO3 + Fe^O3 (^ed.)
38
TOO
60 0
o
500
i
Uoo
Temperal
£
300
200
Figure 29.
NO^ Absorption Curve for K^CO^ + Fe^O^ (bed)
Decoaposed KNO^ + Pe^O^(bed)
Peed - UOO ml/min air
9000 ppn NO
15.85 CO2
_
_ Temperature
Temperature (°C)
_
--- 250 ppm
Figure 30
NO^ Absorption Curve for Decomposed KNO^ + Fe^
(bed)
39
700
6oo £
500
I
300
200
ICC
Figure 31.
NO
x
Absorption Curve for Na CO
j
+ MnO
2
(bed)
KpCO 3 + MnOp (bei)
Feed - 100 ml/air. air
9000 ppE
n:
15.2% CCp
_
_ Temperature
Tempe-
_
50 ppm
100 ppm
Minutes
Figure 32.
NO^ Absorption Curve for KpCO3 + MnOp (bed)
(KOx dropped to less than 40 p p m ) .
Again the mixture with
CO^ ab­
sorbs more than the NagCO^ m i x t u r e .
Unlike the potassium carbonate-ferric oxide m i x t u r e , the manganese
mixture formed clumps which could cause problems in a large scale reac­
tor.
The clumpy material when dissolved in water produced a green
colored solution w h i c h is characteristic of potassium manganate
(KgMnO^) (Partington, 1921; Tyree and Knox, 196l).
A test was per-'
formed in which a mixture of KgCO^ and MnOg was decomposed at a high
temperature in an atmosphere of Kg.
When the decomposition was con­
sidered to have proceeded sufficiently the reactor was cooled down and
the material removed.
This material was hard and initially had a blue
color, indicating that potassium hypomanganate had formed (K^MnO^)
(Partington, 1921).
On being exposed to the room atmosphere, the blue
material gradually turned green and became moist.
The green material
is presumably potassium manganate while the moisture m a y be due to the
formation of potassium hydroxide which then absorbs moisture from the
air.
Potassium permangante ('KMnO^) and a decomposed mixture of sodium
nitrate and manganese dioxide were tested to see if they would absorb
N0x .
Results of these two tests are shown in Figures 33 and 34.
performance of the decomposed nitrate mixture was poor.
The
However, the
potassium permanganate absorbed almost all the NOx in the feed.
The
temperature at which the KMnO^ started absorbing is the temperature
KMnOl
4+ Sloshed)
Feed - 1*00 ml/ain air
9000 ppc NO
15.8? CO^
_
_
__ Temperature
7OC
600
500
too
300
20C
IOC
25 ppm
0
Minutes
Figure 33.
NO^ Absorption Curve for KMnO^ + SiO 2 (bed)
700
600 £
t
500
2
I
Uoo I
300
200
100
Figure 3U.
NO^ Absorption Curve for Decomposed NaNO^ + MnO5 (bed)
k2
at which it decomposes
(l90°C) to potassium manganate (K5M n O ^ ) .
The possibility existed that SiCh in the Ottawa sand reacted with
the carbonate or manganese dioxide and increased their sorbent ability.
It'is known "that SiOg will displace CCL from carbonates at high temp­
eratures to produce silicates.
Figures 35 through 37 are the results
of potassium c a r bonate, sodium carbonate and manganese dioxide tested
independently of each other w i t h Ottawa sand.
The Ottawa sand did not
improve absorption.
Regeneration Studies
Figures
38 and 39 are typical regeneration curves for K CCL-MnCL.
Four absorptions- and regenerations were performed w i t h the K^CO^-MhO^
mixture and there was no indication of loss in sorbent performance.
The concentration of RO
X
obtained on regeneration depends on the flow
rate used,..the amount preabsorbed and the heating rate used in regen­
eration.'
An ROx concentration of greater than 50% was- obtained on
regenerating MhOg-KgCO^ sorbent at a temperature of 600 - JOO0C using
10 ml/min of air.
Furthermore, of the 50%.RO , approximately 49.5%
was ROg and only .5% was. R O .
It is interesting to note the large
amounts of Og that are also given off during decomposition.
This ox­
ygen could be responsible for oxidizing RO to R O g .
Similar regeneration results were obtained with KgCO^-FegO^
mixtures.
After six cycles of regenerating and absorbing, no loss in
8,700 ppo
IIa2CO3 + SiO2 (ted)
Feed - U00 ml/ain air
SWOO ppm IO
700
15.8% CO2
_
_
_ Teeperature
60C
?
Z
500
UOO
5
C
$.
E
300
2,800 ppe
200
100
0
Mlnutea
Figure 35.
NO^ Absorption Curve for Na 2CO 3 + SiO 2 (bed)
I2CC 3 ♦ SlO2 (bed)
Feed - UOO al/aln air
9000 pptn HG
15.8% CO2
_ _
_ Temperature
3,600 ppm
Minutes
Figure 36 . NO^ Absorption Curve for KgCO3 + SiOg (bed)
1+1+
Mo:; ♦ S i O ^ bed)
Feed - UOO m-l/mln air
9000 ppm HO
15.8$ C O 2
_
_
_ Temperature
700
6oo
5,800 ppm
500
UOO
300
200
100
0
Figure 37.
NOx Absorption Curve for MnO 2 + SiO 2 (bed)
5
4$
Regeceration
2 3 ♦ MnO2
I CO
[5.5 gj
Feed — 10 ml/min
--------- Temperature
HO
(H)
t HO.
---------
N0% when NO^% vas equal to 50
Minutes
Figure 38.
NOx Concentration During Regeneration of K 3CO 3 + MnO 3
600 u
Regeneration of absorbent
Minutes
Figure 39 .
Regeneration of K 3CO 3 + MnO 3
k6
sorbent performance was noted.
Other tests on the potassium ferric
oxide mixture indicated that when NO^ is .absorbed at low flow rates
(10 ml/miri) on regeneration, NO is primarily given off, while if ab­
sorption occurred at higher flow rates
given off.
(500 ml/min)? NO^ was primarily
Also after low flow rate absorptions, less oxygen is de­
sorbed than after, h i g h flow rate absorptions.
When desorption was completed a feed of
and NO was passed over
the KgCO^-FegO^ sorbent at temperatures greater than or equal to
TOO0C.
The N
and 0
any NO was present.
content of the exiting stream went up and hardly
This indicates that the NO is reduced b y the sor­
bent to Ng and O g , which corresponds to reports in the literature that
alkali oxides reduce NO to Ng and Og at high temperatures
and M o r e l l o ,
(Harris
1969).
Loading Studies
The time taken for the concentration of NO^ in the exit stream
t o exceed 200 ppm was determined with and without COp in the feed.
r
*
Figures 40 and 4l are typical loading curves.
If COg was present
the loading time was between 30 minutes and 60 minutes.
This corres­
ponds to a conversion of 8.5 to 17% of the .025 moles of sorbent pre­
sent.
When COg was not present the loading time was about equal
to the time needed to theoretically convert all the alkali carbonate
or alkali oxide to an alkali nitrate
(5 - 10 hr).
It was found that
10
Loading Curve
K2CO 3 * Fe 3O 3 (7.5 g)
Feed - LOO ml/min air
Ir Exii
8000 ppm NO
700
15* CO2
— --- - — Temperature
---- H0,
I
0M
Minutes
Figure 40.
Loading Curve for K^CCy + Fe^O^
Loading Curve
K2CO 3 ♦ MnO2 (5.5 g)
Temperature UItO0C
Feed - UOO ml/min air
10,000 ppm NO
15* CO.
Figure Ul.
Loading Cuirve for K^CO^ + MnO,-,
6000
48
if the exit concentration reached 200 ppm when CO^ was present, shut­
ting off the COg would cause the N
concentration to drop below
200 ppm.
COg and H g O .Effects
As was mentioned carbon dioxide affects the absorption of NO and
NO^ but tests showed that no detectable.amounts of COg are absorbed.
The other affect of carbon dioxide is to limit the temperature range
in which absorption occurs.
Addition of carbon dioxide to the feed
will narrow the temperature range from 80 to 20 degrees.
This effect
is less pronounced for the potassium carbonate than for the sodium
carbonate and non-existant for KgMnO^.
'
It is possible by using the free energies of reaction to determine
the equilibrium constants for reaction involving carbon dioxide and
nitric oxide (see Table IX).
The values calculated indicate that the
carbonates do not appreciably decompose by themselves, but by reacting
with NO and O g , decomposition is favored.
When equal amounts (l%) of carbon dioxide and.nitric oxide, along
with 20% oxygen are fed into the reactor, no decrease in sorbent ab­
ility is seen.
It i s n ’t until the concentration of COg is greater than
5% that sorbent performance drops off.
Since the concentrations of
carbon dioxide are far below the values predicted b y equilibrium cons­
tants to cause the reaction to reverse, COg possibly affects an inter-
TABLE' IX:
'"4?/
.Equilibrium’.Constants, .
’for .Beactions’.En-volving .HO .and. COg
Reaction
^
Eree Energy
' Change at 425°C (.Kcal/moleI
Equilibrium
Constant
Ha 2CO 3 = Na2O + COg
52.128
5.09 x IO-17
KgCO3 = K a 2O + COg
68.42
4.3 x IO-22
NagO + 2N0 + 3.5 Og =
Na 2CO 3 +
2NaN0
-143.40
8.58 x IOlllt'
- 91.27
28
3.97 X 10
2N0 + 3.5 Og ? 2NaN03
+ COg
NagFegO, + 2N0 + 3.5 Og =
2NaN0
Na 2CO 3 + Fe 3O 3
47
-150 (estimate)
I x 10
58.24
5.63 x 10
t Fe2O3
Na2F e 2O1^ + COg
-19
50
m e d i a te r e a c t i o n .
F u r th e r m o r e , i f t h e s o r b e n t h a s a b s o r b e d oxygen
p r i o r t o t h e i n t r o d u c t i o n o f a s tr e a m o f NO. an d N ^s t h e n i t r i c
w i l l be a b so rb e d f o r a s h o r t p e rio d o f tim e .
o x id e '
As s o o n a s t h e a b s o r p ­
t i o n o f NO s t o p s t h e i n t r o d u c t i o n o f oxygen t o t h e f e e d w i l l c a u s e
th e a b s o rp tio n t o b e g in a g a in .
, The effect of water on absorption of NO^ was tested b y satur­
ating the air stream at room temperature prior to adding NO to the
stream.
At moisture concentrations of 2-3% no'effect on sorbent per­
formance was detected.
It has been reported that the NO^ absorption
b y the carbonates is facilitated by the presence of.water vapor in the
feed (Lewis, 1975).
DISCUSSION
The potassium and sodium carbonate absorption maximums occurred
at IOO-SOO0C and from SOO-^QO0C.
Generally NOg "was absorbed b y the
carbonates at l o w temperatures (below I O O 0Cj and any NO that was not
converted to NOg passed through the reactor.
The absorption from 100
to 200°C is. possibly due to water being released b y the carbonate.
For
example potassium carbonate had a water content of approximately 8%.
The water is probably due to the decomposition of bicarbonate and the
release of hydrated w a t e r , b oth of which occur in the temperature
range from 100 to S O O 0C.
This would also explain w h y the low temp­
erature absorption m a x i m u m does not occur after the carbonate has
been heated.
It was been reported that the presence of water vapor
improves the absorption ability of carbonates (Harris, Morello and
Peters,
1969).
The high temperature absorption is probably due to a compound
formed during the heating of the carbonate.
Decomposition with lib­
eration of COg and the formation of an oxide is a characteristic prop­
erty of carbonates at h i g h temperatures
(Webb, 1970; pp 304).
The
rate of decomposition depends on the carbonate used and on any imp­
urities present in the material.
The oxide formed depends on its
stability at the temperature in question and the amount of Og present.
It is interesting to note that the reactivity of the carbonates for
absorption of NO^. is in the reverse order of the stability of the
52
carbonates to thermal decomposition.
Tables X and X X list thermal
characteristics of the possible alkali oxides that could be formed.
Thermal studies, on NaO^-NagQ show that the sodium superoxide (NaOg)
decomposes with the rapid evolution of oxygen up to 280°C leaving only
sodium peroxide (Na^0^)
(Webb, 1970; pp 242).
Decomposition of Na^Og
into NagO and oxygen does not take place until the temperature is
above melting (460°C).
In decomposing sodium carbonate in the reac­
tor the compound formed had a yellow color indicating possibly that
Na^Og was formed.
Common oxides of potassium are potassium superoxide (KOg) and
potassium peroxide (KgOg); the superoxide is the most common.
When
KOg is heated, reversible decomposition into oxygen and a mixture of
KgOg and KOg takes place at 300 to 400°C (Yost and Russell, 1944).
A
yellow substance was formed when potassium carbonate was heated in
the reactor indicating that the oxide'formed was K O g .
It is possible then, that the active species in absorption at
higher temperatures is not the alkali carbonate but rather the alkali
oxide w h i c h is formed from the carbonate.. This would explain why ab­
sorption increases after the carbonate has been heated to high temp­
eratures.
■TABLE X:
M e lt in g -and --BoTBing-BoLnts-Lof ' A l k a l i . C arB o n ates- a n d O x id e s*
.Formula
Formula
weight
Bip(0 C)
LigCOg
73.89
735
Color
L p ( 0 C)
white
'3590
I
Ha 2CO 3
KgCO3
105.99
854
d 600
white
-138.21
896
d670
white
LigO
29.88
1570
LigOg
1:5.88
dl'6o
NaOg
54.99
d 280
Na2O
61.98
N a 2 Og
77.98
KDg
71.1
K2 O
94.20
d350
d88l
110.20
490
d
.
%
*(jang. 1967; Samsonov, 1973)
d!93
white
920
d 46o
• 402(380)
d657
yellow
d
yellow
white
54
Table
Ll2O
XE
-
--Thermal Stability}:' Characteristics -of Alkali Oxides*
Thermally stable.
Volatility 'increases .and thermal stability
decreases in the series Li— C s .
Li2O^
Decomposes Trith- liberation of oxygen at l 60oC.
Decomposes
completely at 315'-343°C.
Na2O
Volatile above 1300°C.
NagO2
Some loss of active oxygen is observed on heating at 311~400°C
vigorous decomposition begins at 540°C; gives up active oxygen
completely above
6j^°C.
NaOg
Decomposes on heating;
decomposes completely above 540°C.
K 2O
Decomposes at 300-400oC without melting to metallic patassium
and peroxide.
Melts without decomposition.
Thermal stability of peroxides
rises from K to Cs.
KO2
Decomposes completely at 5U3°C.
* (Samsonov, 19731
55
Effect o f '6xid.es on the Decomposition of Carbonates
As was discussed earlier the effect of impurities in the carbon­
ate can m a r kedly affect its decomposition rate.
the results of CO
Shown in Figure 42 are
released from mixtures of carbonate and metal oxide
when decomposed in an air stream.
The presence of the metal oxide in­
creases the rate of decomposition of sodium and potassium carbonate
considerably.
This increase in CO^ liberation m a y be responsible for
the increased sorption ability of the carbonate due to the production
of active alkali oxides.
oxygen or E
It was also found that the presence of
in the feed increases the rate at which COg is released
from the carbonate.
A test was performed using iron powder and sodium
carbonate as the sorbent, although the thermal conductivity of this
mixture would have been higher than that of the other metal oxideNa 0CO 0 mixtures, no increase in NO
absorption occurred above that of
just NagCOg.
The alkali oxide formed from decomposition of a Na^CO^ or NaNO^
is reactive and would combine with Fe^O^ to form a ferrite or ferrate
(NagO-FegO^).
The ferrates are known to be strong oxidizing agents
and it is possible that if a nitrite was formed it would be oxidized
to a nitrate.
The possible reactions that may be occurring are as
f o l lows:
NagOFegO
+ 2N0 + 3/2 Og = 2NaN0^ + FegO 3
N a 2OFe 2O 3 + 2N0g + 1/2 Og =
ENaNO3,+ Fe 3O 3
■
56
Decomposition
Peed - 500 ml/min air
+ Fe,
♦
HnO,
0
Na CO.
l2vv3
Figure
h2
Effect of Metal Oxides on Decomposition of Carbonates
57
It is also possible that NagO2FegQ^ is .formed and undergoes the same
reactions except a half a mole less of oxygen would be needed.
reactions are listed in Table XIX.
Other
In a test using only nitrogen and
NO the ferrate from the nitrate reduced the NO concentration from a
level of
7500 ppm to 3500 ppm, indicating that a reaction will occur
without oxygen present in the feed gas.
The temperature at which absorption occurs and the activity of the
carbonate mixtures compared to those prepared from the.' nitrates might
be explained b y the following paragraph taken from Arvid Eedvall *s
Solid State C h e mistry, "It is in many cases advantageous to ‘start, not
from original oxides but from structurally defect and consequently
particularly active oxides which are formed by thermal decomposition.
An example is the production of ferrites and chromites b y homogenised
heating of the required oxide w i t h iron carbonate or of magnesium
carbonate with chromium oxide, respectively.
In powder mixtures of
ferric oxide and oxides of alkaline-earth metals, the preparations with
structural defects formed ferrites far more rapidly and completely
than the preparations without defects. The differences only disappear
at temperatures above 900°C as the irreversible defects vanish.
Innumerable experiments involving various types of changes -of
structure have provided a basis for a general rule which states that
the temperature (or temperature range) where new phases are formed b y
transformation coincide with a relative maximum of reactivity in a
substance."
Temperatures at w h i c h the nitrates undergo a change are listed in
Table XIII.
Sodium nitrate and potassium nitrate decompose in the
temperature range at w h i c h NO^ is absorbed by the sorbents.
Chemical
analysis also indicates that nitrates are formed when NO^ is absorbed.
An alternate method b y w h i c h the reaction could proceed is
through an oxidation step on the metal oxide.
Before discussing this
Table XIX.
-Possible Reactions of the Sorbents
AHgp8Ckcal/mole)
Reaction
Na 5CO 3 = Na2O + CO2
76.8
K 2C O 3 = K 2O + CO2
9^.05
N a 2CO 3 + 1 / 2 -O2 = N a 2O2 + CO2
55.6
K2CO3 + 1/2 O2 = K2O2 + CO2
62.U5
2NaN03 = NagO + 2N0 + 3/2 Og
166.88
2KN0
192.29
2NaN0
=
KgO + 2N0 + 3/2 Og
= Na2O2 + 2N0 + Og
1^5.68
160.69
2KN03 = K2 O2 + 2N0 + O2
N a 2CO 3 + 2N0 + 3/2 O2 = 2NaN03 + COg
K 2 CO 3 + 2N0 + 3/2 O2 =
2KN0
+ COg
-
90.08
-98.27
59
Table XII; Continued:
Reaction
A Hggg (.kcal/mole)
NO + 1/2 O2 = RO2
-13,51
N a 2O + 2N02 + 1/2 Og = 2NaN0^ .
-139.86
KgO + 2N02 + 1/2 O2 = 2KNO 3
-166.30
2NaN03 + Fe 2O 3 = Na2F e 2O^ + SNOg + 1/2 O2
176.67
N a 2CO 3 + Fe 2O 3 = N a 2Fe 2O 4 ■+ CO2 •
121.75
(32.06)*
N a 2Fe 2O 4 + 2N0 + 3/2 O2 = 2NaN0
+ T e 2O 3
K 2 CO 3 + M n O 2 + 1/2 O2 = K 2M nO 4 + CO2
Na 2MnO 4 + 2N0 + O2 = 2NaN0
+ MnOg
K 2MnO 4 + 2N0 + O2 = 2K N O 3 + MnO2
- 211.78
221.98
-116.78
-133.22
* Matsui., Mototaro and Sakamaki., Tomonosuke, J. S o c , C h e m . Ind. (japan)
30, 29-36 C1927)
6o
0
TABLE- X E H . Melting and Decomposition. Temperatures of. Alkali Nitrates,
■ ■
•'
;•' '"Nitrites and Hydroxides -(Jang, .196.7) "(CRC"H A n d L o o k '4|?th ed)
Eormula
Weight
-DipC0 C I
LiNO 3 .
68.95'
25U
d 600
NaNO0
85. Ol
310
d 300
ENO
1 0 1 .1 0
337
d3U0
LiNO2
52.95
100
d l8 5
NaNO2
69.00
271
d320
KNO2
8 5 .ll
UUo
d350
LiOE
23.95
U62
NaOH
U 0 .0 0
318
KOH
56.11
360
'3
O
Compound
'61
mechanism the oxides used in the oxidation of CO to COg will "be des­
cribed,
The catalytic activity series of metal oxides for the oxid­
ation of CO is from the most active to the least active (Dixon and
, ,
Lo n g f i e l d , i 960):
Co
+2
,Cu
+1
,Ni
+2
,Mn
Cr+ ^, Th+^ , Zr+^ , V+ ^ , Hg+ ^, and Al+ ^.
+Ii
, Cu
+2
, Fe
+ ’R
, Zn
+2
, Ce
+U
,
There is also a qualitative
connection between the activity of the oxide and their semicond­
ucting properties.
The most active oxides are the P type which include
manganese dioxide (Ashmore, 1967 ; pp 239).
In the P type oxides,
electrical conduction occurs b y changes of charge or motion of pos­
itive holes.
N type oxides
(eg. FegO^) which are the next most ef­
fective oxidizers of CO, conduct electricity by free electrons.
At high temperatures carbon monoxide attacks the surfaces of P
and N type oxides to form carbonate ions.
The proposed mechanism is
one in which oxygen is absorbed as O'. •and CO reacts with two 0
to
form the carbonate, which decomposes to COg (Ashmore, 1967; Thomas
and Thomas, 1967 ).
A similar scheme can be thought to occur for nit­
ric oxide, since NO and CO are very similar molecules.
The NO^
formed when NO is absorbed would decompose to produce NOg wh i c h would
react w i t h the alkali carbonate or oxide to form a nitrate.
Absorp­
tion of NOg at low temperatures does occur for the carbonates and
studies indicate that NOg is absorbed as NO^ ions on AlgO^ and NiO
surfaces (K h o z i n g e r , 1976) .
Pt is also known that NO will displace
CO from the surface of metal o x ides, indicating that the NO molecule
62
H a s a strong affinity for metal oxide surfaces.
Release of Oxygen from Potassium Carbonate and Manganese Dioxide or .
Ferric Oxide M i x t u r e s .....................
................... ...... •
Tests were performed to determine if oxygen was given off during
decomposition of KgCO^ + MhO^ or K^CO^ + Fe^O^.
As is shown in
Figure 43 and 44, the manganese dioxide mixture releases O^ when de­
composed, while the ferric oxide mixture does not.
The decrease in'
Ng at 200 - 250°C is not due to Ng being absorbed, but from the de­
composition of bicarbonate present to give off HgO and C O g .
Decomposition of manganese dioxide to MhgO^ and Og is responsible
for the large Og p e a k around 450 - 550°C.
Integration of the Og peak
indicates that approximately 1.5$ of the original
ganese dioxide releases O g .
.025 mole of man­
It is possible that MhgO^ is an active
species in the absorption process if it doesn't" revert back to MnOg
below 450°C in the presence of O g .
Decomposition of the carbonate appears to reach a final value for
the MnOg-KgCOg mixture while the FegOg-KgCOg decomposition continues
for a longer period of time as indicated b y the nitrogen composition
of the exit stream.
the original
An approximate analysis indicated that 11$ of
.025 mole of carbonate present in both mixtures releases
COg and the total amount of carbon dioxide released after 70 minutes
is the same for b o t h mixtures.
63
Decomposition of
600 u
Minutes
Figure 43.
Amount of
Released From K 5CO 3 + MnO 5
Decomposition of
•6 0 0
Feed - 10 ml/min I3
Figure 44.
Amount of O5 Released From K5CO3 + Fe5O3
0
A
TIie possible reactions of the manganese mixturein,' an
atmos­
phere are as follows:
1)
K 2CO 3 = K2O + CO2
2)
GMnO2 = M n 3O 3 + 1/2 O2
3)
K 2CO 3 + M n O 2 = K 2MnO 3 + CO2
4)
K 2CO 3 + M n O g = K 2MnO^ + C O
5)
K 2CO 3 + GMnO2 = GK 3M n O 1^ + GCO2 + CO
,
All of the reactions are endothermic and would be favorable at higher
temperatures.
Equilibrium calculations are not app]icable since the
gaseous products are removed by the nitrogen stream but the rates of
the reaction may play an important role. .Exactly to what extent each
reaction takes place is hard to determine.
or K 3MnO^ has been suggested earlier.
The formation of KgMnO^
The presence of
stream indicates that reaction 2 takes place.
in the exit
Analysis of CO was not
possible at the time so definite conclusions can not be drawn on the
extent of any reactions.
Similarly, the reactions that take place between KgCO 3 and Fe 2O 3
c o u l d n ’t be positively identified but some of the likely ones that
could take place in the inert Eg atmosphere are:
1)
K 2CO 3 = KgO + CO2
2)
K 2CO 3 = K2 Q2 + CO
3)
K 2CO 3 + Fe 2O 3 = K 2OFe 2O3 + CO2
4)
K 2CO 3 + Fe 2O 3 = K 2O 2Fe 2O 3 + CO
65
Absorption of CL b y Decomposed Potassium Carbonate, Ferric Oxide and
_______ ' ' '
__________
Manganese Dioxide Mixtures
After decomposing tb_e carbonate metal oxide mixtures in a
Ng
stream, the feed stream was switched over to air and the amount of
absorbed recorded (see Figure U 5).
The initial large absorption of '
Og b y b o t h mixtures is probably caused b y the already existing com­
pound formed when the carbonate decomposed and reacted w i t h the metal
oxide.
Since the manganese mixture initially absorbs more Og and
d o e s n ’t give off C O g , the carbonate in this mixture is largely de­
composed.
Furthermore, once loaded, the manganate does not absorb Og
until.the temperature drops to below 600°C which corresponds to the
reaction MhgO^ + 1/2 Og = 2Mn0g.
The absorption of Og between 600 -
450°C was reversible in that Og
was given off above this temperature
and absorbed when the temperature was lowered.
On the other hand,
the ferric oxide-carbonate mixture continually absorbs oxygen and
gives off carbon dioxide indicating that the carbonate i s n ’t fully
decomposed.
Then when the temperature drops, the carbonate no longer
decomposes and subsequently Og absorption stops.
These tests indicate
that the oxygen is absorbed, therefore MD may be reacting with the ox­
ygen as an intermediate step in the absorption process.
66
Absorption of
Feed — 10 ml/ialn air
•600
0
Minutes
Figure 45.
Absorption of
by K^CO^ + MnO 5 and K^CO^ + Fe 5O^
APPLICATION
■ It was estimated that approximately 1.25 million standard cubic
feet per minute of gas containing 1.5% NO VrillxIiave to be handled for
a 1000 Megawatt MHD generator.
to be 2.37 x 10
k-
gmole/min.
laboratory reactor,
In terms of moles of NO this comes out
Under the test conditions used in the •
.025 mole of sorbent was able to absorb 1.63 x IO-
mole/min of NO for 30 minutes before having to be regenerated.
If the
same conditions are applied to the MEED plant then 3.64 x 10^ moles of
sorbent would be required every 30 minutes.
Assuming a continuous
p r o c e s s , taking a one hour cycle for completion, the total amount of
sorbent that would have to be purchased is 7.28 x 10^ moles or approx­
imately 40 tons.
The initial costs of sorbent mixtures that could be
used for absorption of NO
is shown in Table XIV (individual costs for
chemicals are listed in Table X V ) .
TABLE XIV.
Mixture
In the study performed to deter-
Costs of 7.28 x IO^ Moles of Sorbent Mixtures
Cost .in' Thousands of Dollars.
KgCO^-FegOg
520
NagCOg-FegOg
165
N a 2COg-MnOg
320
KgCOg-MnOg
•
673
mine the economic feasibility of producing nitric acid from a 700 Mega­
watt MHD generator, silica gel was used as the sorbent and oxidizer at
68
cost of over
2 million dollars Offright and Youderion, 1976).
Even
with, this h i g h expenditure for a sorbent the plant showed a projected
rate of return of 31$.
Scaling up the cost of the silica gel for a
1000 Megawatt generator indicates an investment of over 3 million
dollars.
Comparing this cost to the cost of the solid sorbents shows
that the solid systems could save over 2.5 million dollars.
Further­
more, the optimum ratio for the sorbent mixture has not been deter­
mined nor has the maximum flow rate, both of which would further
reduce the cost.
69
' 'TABLE XV. ' 'Costs:'of 'CMrfttcal^ ■ ■
CIiemLcal
NSgCOg
dollar s'/lB
028
dollars/IL mole
2.97
KgCOg
.20
25.00
FegOg
'.QhG
7.36
MnOg
.21
17.00
KMnOj^
CO
82.00
BaNOg
065
KEOg
KCl
CaCOg
CaClg
5.53
.10
10.11
..02
1 .U8
.01
1.00
' .035
3.85
EiO
2.00
Ih9.h2
COgOg
4.27
■708.22
FbOg
.66
157.74
AlgO
.16
16.32
CONCLUSIONS
1.
Ferric oxide or manganese dioxide m i x e d on a one to one mole "basis
with, potassium or sodium carbonate will absorb KCf. in a temp­
erature range from 380 to U50°C.
U s i n g these absorbents, the
concentration can be reduced from
10,000 ppm to less than 200 ppm
from a g a s containing 1 # NO, 15$ C O g , 17.5$ Og and 66.5$ Ng at flow
rates of 500 ml/min.
The effective operating time of .025 moles
of sorbent mixture was' between
2.
.5 and I hour.
Decomposed mixtu r e s of sodium nitrate and ferric oxide (ferrates)
are effective NO^ absorbers when the ratio of Og to NO in the feed
is greater than 3:1.
3.
Nickel o f "cobalt oxides; mixed w i t h sodium carbonate show potential
for absorbing N Q x at high temperatures.
4.
The order of the solid sorbents for absorbing NO is as follows
(in order of increasing ability):
LiCO^, NagCO ^5 K g C O ^ , de­
composed NaNO^ + F e g O ^ , NagCO^.+ FbgO^, KgCO^ + FegO^, NagCOg +
M n O g , KgCO
5.
+ M n O g , KMnO^.
Carbon dioxide at concentrations of greater than 5$ reduced the
effective operating times of sorbents from several hours to less
than one hour.
KgCOg mixtures, are less affected b y the C Og.
6. Water vapor at concentrations of 2-3$ does not affect the gprption
ability of the solid sorbents.
Tl
7.
Alkali' nitrate?:' are formed %&en NXy i?, absorbed b y alkali, carb­
onate— ferric oxide or manganese dioxide mixtures-.
8 . : Nitrogen dioxide is.- absorbed at l o w temperatures b y the carbonate
metal oxide mixture. .
9.
.
i
I-
Nitric oxide is not absorbed at room temperature but is absorbed
,at higher temperatures C^QO0C I b y the-carbonate-metal oxide sor­
bents.
10.
Nitrogen dioxide is- formed when sorbents are regenerated at temp­
eratures between 6>0Q and TOO 0C.
11.
Formation of NO^ during regeneration of ferric oxide-carbonate
sorbent is affected b y the flow rates used during absorption of
NO^.
12.
Higher flow rates favored the formation of N O g .
Carbon dioxide is- not absorbed in any detectable amounts b y the
sorbents.
13.
At h i g h temperatures
CTOO0C I in an inert atmosphere, NO is red­
uced to N g and Og b y potassium carbonate-ferric oxide mixtures.
lit.
The carbonate-oxide mixtures appear to be more economical than
silica gel for the absorption of NO^ from the MHD exhaust.
,
RECOJWEKDATIDITg'
The research, performed has- ihdi'cated that carhonatepoxide m i x ­
tures- are effective U
serhents'.
Data is- n o w required to better
evaluate the performance of each, sorhent mixture.
Therefore, the
following factors need to Be k n o w n ;
1.
Optimum f l o w rates, temperatures and stoichiometric ratios
for sorhent mixtures',
2.
Ra t e s and mechanism of reaction for absorption and regener­
ation.
3.
Effect of. different concentration of C O ^ 5
and NO.
U.
Surface area-of sorbents.
5.
Heats- of reaction for regeneration and absorption.
Since there is evidence indicating that oxidative properties of
the metal oxides are important, mixtures known as Hopcalites
w i t h metal oxides] should be tested w i t h the carbonates.
(MhO^
In conjunc­
tion w i t h the oxidative properties, the semiconductive properties
should be evaluated to help determine a mechanism.
Further investigations in the reduction of NO at temperatures
greater than T O O 0C b y the sorbents m a y also be valuable.
Application
of ..the. sorbents for automobile-emissions.- control should- be investigated
LIST OF REFERENCES
A s h m o r e , P. G.
1963.''C a t a l y s i s 'a n d 'Inh i b i t i o n 'o f 'Chemical Reactions.
P. 239.
Butterworth. Inc.
Bartok, ¥. , Crawford, A. R., and S k o p p , A.
1971.
Control o f 'NO
Emissions''f r o m 'Stationary Sources. ' CIiem. Eng. Prog.
X
67
C2):
64-72.
'
Ch-emical Marketing Reports.
C h i l t o n , T.
BE.
............
Aug. 9, 1976.
1968. ' Strong Water.
Schnell Publishing Co. Inc.
89. Mass M..I.T. Press.
P. 53,
D i x o n , J. K., and Lo n g f i e l d , <T. E.
i 960. Catalysis.
N e w York:
Reinhold Publishing Corporation.
VII: 305
D u r r a n t , J. P. and D u r r a n t , B.
1962.
Introduction to Advanced
Inorganic Chemistry.
P. 1012, 1013, 1014.
New York:
John
W i l e y & Sons Inc.
Harris, S. W., M o r e l l o , E. F., and Peters, G. H.
3, 459, 494.■ Aug. 5.
1969.
U..S. Patent
H a l s , F. and Jackson, W. D.
1969. MHD Power Generation - Economic
and Environmental Implication, Tenth Symposium on Engineering
Aspects of MHD.
M.I.T. Cambridge Massachusetts.
Hals, F. A. and Lewis, P. F.
1973.
Control Techniques for Nitrogen
Oxides i n MHD Power Plan t s . N ew York:
ASME Publication.
H e d v a l l , A. J. 1966. Solid State Chemistry.
Publishing Company.
Jang, G. J.
1967.
Molten Salts Handbook.
P. 25-27.
New York:
K h o z i n g e r , H.
1976.
Advances in Catalysis.
Academic Press, Inc.
Elsevier
Academic Press.
25: 230-232.
N ew York:
Koval, E. J., and Roberts, M. A.
i 960. How Does Nitric Oxide Affect
Reactions of Aqueous Nitrogen Dioxide?
I n d . E n g . 'Chem.
52: 1011.
Leung, L., and Gidaspow, D. 1974. Iron Oxide Sorbents for
Regenerative Sorption of N0_^. ''AICHE J . 20(.5): 840.
Sept.
Lewis,
H.
1975. ' M t r o g e n 'Oxides Removal.
Corporation.
New Jersey:
Noyes Data
Partington, J. R.
1921.
A Text-Book of Inorganic Chemistry.
N e w York:
St. Martins Press, Inc.
Pepper, J. W. , E u s t i s , R. H. and K r u g e r , C. H. 1972.
NO Concentration
in MHD Steam Power Plant Systems.
Twelfth. Symposium
Engineering A s p e c t s 'of'Magnetohydrodynamics. Argonne
National Laboratory, Illinois.
Pollution Control Technology.
Samsonov, G. V. Ed.
1973.
Seinfeld, J. K.
1975.
'Fundamentals..
1973.
Research, and Education Assoc.
Oxide Handbook.
IFl/PLENUM.
Air Pollution Physical and Chemical
McGraw-Hill Inc.
T h o m a s , J. M., and T h o m a s , J. ¥.
1967.
Introduction to the Principles
of Heterogeneous C ata l y s i s . . P. 367-375.
New York:
Academic
Press.
Tyree, Y. S. J r . , and Knox, K.
196l.
Textbook, of Inorganic Chemistry.
P. 235.
N e w York:
MacMillan Company.
Walters and Goodwin.
197^.
Federal Emission Standard for Stationary
Sources.
AICHE Symposium No. 1 3 7 « 70:
3.
Webb, T. L.
1970.
Differential Thermal Analysis.
Academic Press, Inc.
I:
2k2.
New York:
•Wright, S., and Y o u d e r i o n , R.
1976.
Economic Analysis of Nitric Oxide
and Ammonium Nitrate Production Facilities on MHD Exhaust.
Montana State University.
June 4. Dept, of C h e m . Engineering.
Yost, D. M., and Russell, H. Jr.
±9kk. -Systematic Inorganic Chemistry.
■ P. .385-386.
N e w York:
Prentice Hall, Inc.
MONTANA STATE UNTVrBerrv i
3 1762 10013635 5
N378
Ev64l
cop.2
DATE
Evonixik, Christopher J
Solid sorbents for the
recovery of nigrogen
oxides ...
ISSUED TO
^
& 6//
-X